141 results about "Mixed model" patented technology

The invention discloses an automobile seat framework machining defect detection method based on machine vision, and particularly relates to the technical field of defect detection. The method comprises the following steps: constructing a multi-angle image acquisition and edge reflection modeling module aiming at complex defects such as weld joint pseudo soldering, microcracks, hole site deviation and collapse deformation, extracting weld joint continuity, edge integrity and hole site geometric consistency characteristics by using a deep neural network, and generating a structural feature vector; defect type recognition and credibility scoring are completed through small sample anomaly modeling and Gaussian mixture model classification, sub-pixel-level coordinate labeling of defect positions is achieved in combination with a Gaussian fitting algorithm, a defect positioning map is output, traceability analysis and severity grading are conducted based on historical data comparison, and the defect positioning accuracy is improved. The method is suitable for industrial online detection and quality closed-loop control.

The invention discloses an industrial equipment defect detection method based on multi-modal feature fusion and dynamic optimization, and belongs to the technical field of computer vision, and the method comprises the steps: 1, multi-modal industrial data collection: deploying multiple sensors in a production line, collecting data in multiple periods, constructing a defect-free and multi-type defect sample library, and carrying out multi-modal industrial data collection; a time-space aligned multi-modal label is marked; step 2, data enhancement and defect synthesis; step 3, multi-modal hybrid model training: constructing a hybrid network, and performing pre-training and fine tuning by using a dynamic loss function; step 4, edge end dynamic optimization and deployment: edge end reasoning is realized through dynamic knowledge distillation, and model fine tuning is automatically triggered when false detection and missing detection are found; and step 5, intelligent labeling and result visualization: a front-end interface displays a detection result in real time. The problem that a current target detection framework is not high in small target recognition accuracy and low in efficiency is solved, and the reliability of industrial equipment defect detection is improved.

Methods and systems implement a pleno-generation face video compression framework with bandwidth intelligence for generative models and compression. Heterogeneous-granularity facial description regularizes long-term dependencies between video frames and compensates for motion estimation errors caused by compact representations of motion information. A generative decoder reconstructs heterogeneous-granularity visual representations, providing auxiliary visual signals for attention-based recalibration of a GFVC-reconstructed face signal. A coarse-to-fine generation strategy avoids error accumulation. High efficiency for heterogeneous-granularity signal compression is achieved by two different entropy-based signal compression methods: heterogeneous-granularities feature representation from the key-reference frame as hyperpriors to optimize the entropy model for compressing heterogeneous-granularity feature from subsequent inter frames, and a feature difference operation for heterogeneous-granularities feature representation between key-reference and subsequent inter frames, such that the entropy model only compresses heterogeneous-granularities feature residual for redundancy reduction. Mixed-model dataset generation and training and model-specific dataset generation and training are also provided.

The invention discloses an interlayer gold purity evaluation method and device based on a hybrid model, equipment and a medium. The method comprises the following steps: collecting original data; inputting the processed original data into a hybrid model for training to obtain a trained hybrid model; wherein the hybrid model comprises a feature extraction module, a weighted fusion unit and a classification module which are connected in sequence; collecting a second pulsed eddy current signal and a second supplementary feature corresponding to the gold to be detected; inputting the second pulse eddy current signal and the second supplementary feature into a trained hybrid model, and outputting a prediction vector; and converting the prediction vector into a one-hot code, and obtaining the number of the corresponding gold to be tested through the one-hot code. The problem that an existing classification mixing model cannot be suitable for a highly-adulterated gold purity detection scene is solved, the classification precision of adulterated gold purity detection is improved, and the application range of the classification mixing model in the field of scarce sample detection is widened.

The invention relates to a meteorological-based dynamic graph network photovoltaic power station group ultra-short-term prediction method, which is characterized in that a dynamic space-time graph network is constructed, photovoltaic power stations in a region are regarded as a complex network system connected by meteorological fluctuations, each power station is taken as a node, and the propagation relationship of the meteorological fluctuations among the power stations is represented by edges of a graph. The weight and time delay of the edge are dynamically adjusted according to real-time meteorological data, and dynamic interaction between power stations, the overall trend under the stable meteorological condition and rapid fluctuation caused by sudden weather events are accurately captured in combination with a mixed model of a graph neural network and a recurrent neural network. Compared with a traditional method, the method has the advantages that the prediction precision, the real-time response capability and the physical interpretability are remarkably improved, scattered photovoltaic power stations are integrated into a mutually associated dynamic system, an efficient and reliable solution is provided for power prediction of a power station group, and the operation stability and the power grid dispatching efficiency of a large-scale photovoltaic system are enhanced.

The invention provides a mobile formwork intelligent monitoring method and system based on digital twinning and multi-modal perception, and belongs to the technical field of building construction intelligence, and the method comprises the steps: building a multi-modal sensor network on a mobile formwork entity; establishing a digital twin platform: constructing a BIM model mapped with a mobile formwork entity, pre-burying a virtual monitoring point in the BIM model, performing matching calibration with a sensor of the multi-modal sensor network, and fusing multi-source sensor data acquired by the multi-modal sensor network in real time; calculating a construction health index (SHI) in real time based on the multi-source sensor data; on the basis of the multi-source sensor data, the instability probability is predicted in real time through an AE-LSTM mixed model; meanwhile, predicting a weight factor of a construction health index SHI; and according to the SHI and the instability probability, jointly evaluating whether a mobile formwork adjustment instruction is generated or not and sending out an early warning. The safety state of the movable formwork is controlled in real time, and the problem that the instability trend cannot be monitored in real time is solved.

The invention discloses a medium and low voltage line abnormity identification method based on a deep neural network, and the method achieves the comprehensive coverage of key nodes of a power distribution network through the four-stage deployment of station-line-transformer-household, completely captures the abnormal signals at the tail end of a branch line, provides sufficient data support for deep feature learning, solves a problem of a conventional monitoring blind area, and improves the recognition precision of the power distribution network. A convolutional neural network (CNN) and long short-term memory (LSTM) network mixed model automatically extracts time-frequency domain joint features, manual feature design is not needed, subtle changes of transient waveforms are effectively captured, the anomaly identification precision is improved, the limitation that a traditional method depends on artificial experience is broken through, potential fault hidden dangers can be found in advance, sufficient processing time is won for operation and maintenance personnel, and the operation and maintenance efficiency is improved. Through rapid positioning, the fault processing period is shortened, the power failure time is reduced, the power supply reliability is improved, fuzzy C-means clustering is combined with topology analysis, accurate positioning of abnormal sections is achieved, the manual inspection workload is reduced, the fault processing efficiency is improved, and the operation and maintenance cost is reduced.

The invention discloses a main beam formwork erection elevation adjustment method based on machine learning, and belongs to the technical field of elevation adjustment. The method comprises the steps that a space-time database is constructed by collecting data of the construction period of a cable-free section; processing the time sequence data by adopting an LSTM-Transform hybrid model, extracting long-period characteristics, and outputting an initial prediction value of the elevation; analyzing nonlinear variables such as a cable force change rate, a sunlight gradient and a material age by using an XG-Boost algorithm, and outputting a compensation factor; fusing and generating a joint prediction value; constructing a reinforcement learning agent, and outputting an adjustment instruction by taking a construction stage as a state space, taking an elevation adjustment amount as an action space, minimizing deviation between a predicted value and a measured value and taking construction stability as a reward function; the formwork erecting elevation is adjusted according to the driving hydraulic system, and the database is updated in real time. According to the invention, through multi-dimensional data fusion and intelligent optimization, the elevation adjustment precision and stability are improved.

The invention discloses an urban space-time evolution analysis method based on multi-source historical data. The method comprises the following steps: 1) acquiring and digitally processing historical city maps, point-of-interest data and realistic paintings in multiple periods; 2) performing syntactic analysis on the urban road network and the water system by adopting a mixed model of topology and angle distance, and calculating core indexes such as integration degree and angle selection degree; 3) based on an optimization algorithm, carrying out normalization processing on the spatial syntactic index and the POI data; 4) extracting dynamic city information from the historical drawing; 5) performing statistical test on dynamic data extracted by drawing and a spatial syntactic analysis result; and 6) analyzing the space-time aggregation mode and function evolution of the public service facility by adopting a location quotient formula and kernel density estimation.By constructing a closed-loop research framework of multi-source data fusion, algorithm optimization and quantitative verification, a highly reliable scientific method is provided for deeply revealing an internal driving mechanism of urban form evolution.

The invention relates to the technical field of electronics, and discloses an urban river pollutant traceability method and device, electronic equipment, a medium and a product. The method comprises the steps that when it is monitored that target pollutants of any target monitoring section exceed the standard, the target monitoring section serves as an end point; reversely simulating an upstream diffusion path of the target pollutants by using a pre-constructed mechanism and data hybrid model to obtain theoretical concentrations of the target pollutants in a plurality of upstream river sections; determining at least one traceable river reach in the plurality of upstream river reaches based on the theoretical concentration of the target pollutants; pollution characteristic data of a water body in the traceable river reach is obtained; performing similarity calculation on the pollution characteristic data and characteristic data of a plurality of pollution sources in a pre-constructed pollution source characteristic knowledge graph to obtain a similarity value between the pollution characteristic data and each pollution source; and determining at least one target pollution source in the plurality of pollution sources based on the similarity value, thereby realizing dual precision of qualitative and quantitative analysis of the pollution sources, and greatly improving the tracing efficiency.

The invention relates to the technical field of machine vision and optical measurement mechanics, and discloses a deformation identification and measurement method based on machine vision and deep learning, and the method comprises the following steps: S1, system calibration and base library construction; s2, offline training of the hybrid model; s3, real-time decoupling and coefficient regression are carried out; and S4, final assembly and physical quantity calculation of the deformation field. According to the method, a deformation field is decoupled into a global nonlinear deformation field () and a residual deformation field () through S3, and the global nonlinear deformation field () and the residual deformation field () are superposed. The physical significance and the stability of a deformation main body are ensured by utilizing the physical deformation base library constructed in S1 and model reconstruction; meanwhile, a luminosity residual image () generated in the S3 is used for driving module reconstruction, local high-frequency disturbance which is not covered by a physical model is accurately compensated, a high-dimensional deformation field is decoupled into low-dimensional global physical prior regression and residual correction, and efficient real-time measurement is achieved; meanwhile, numerical difference is replaced by analytic derivation, so that the signal-to-noise ratio and the measurement precision of the strain field are remarkably improved.

The invention provides a real-time time sequence prediction system and method based on a dynamic weight hybrid model, and belongs to the technical field of time sequence prediction and machine learning. The prediction system comprises a data generation module, a real-time data caching module, a model initialization module, a prediction model selection module, a single model prediction module, a mixed model prediction module, a result display module, an error calculation module, a dynamic weight optimization module and a reset module. According to the method, a dynamic weight fusion strategy is adopted, and contribution weights of ARIMA and LSTM models are dynamically adjusted according to real-time data characteristics (such as data stability, non-linear degree and fluctuation amplitude). A dynamic weight mechanism solves the problem of'one-cut 'of a fixed weight hybrid model, so that the model can maintain the optimal performance in a linear stable scene (such as a new energy output stable time period) and a nonlinear fluctuation scene (such as an extreme weather time period), and the generalization ability of prediction is remarkably improved.

The application discloses a full waveform decomposition method with an oscillation signal in an airborne laser radar, belongs to the technical field of laser radar data processing, and is used for waveform decomposition and comprises the following steps: reading and converting original data, performing data blocking, selecting a wavelet threshold for denoising of each block of data, constructing a mixed model, selecting initial parameters of the mixed model, performing particle swarm optimization, obtaining optimized parameters, constructing output data of each block of data, splicing all the output data, and completing full waveform decomposition. The application removes almost all high-frequency noises, has a good retention degree for signals, directly performs full waveform decomposition, fits an oscillation signal, reduces the influence caused by the oscillation signal, and extracts distance data of the airborne laser radar. The multi-point search and randomness of the particle swarm make it more possible to find a global optimal solution, the particle updating is naturally parallelized, is suitable for distributed computing, and the calculation speed is accelerated.

The invention relates to the technical field of water pollution treatment, and discloses a method and a system for dynamically adding a complex microbial inoculant in combination with water quality monitoring. Real-time water quality parameters are collected, and the real-time water quality data are preprocessed to obtain preprocessed water quality data; establishing a mixed model comprising a water quality prediction branch and a fungicide adding branch; inputting the preprocessed water quality data into the mixed model, and determining the dosage of the complex microbial inoculant required at the current moment through a microbial inoculant adding branch to obtain a preliminary dosage; obtaining a water quality parameter prediction value in the future 1-2 hours through a water quality prediction branch, and dynamically adjusting the initial dosage by adopting an MO-IPSO algorithm based on the water quality parameter prediction value to obtain an optimized dosage; sending the optimized adding amount to adding equipment so as to control the adding equipment to add the complex microbial inoculant according to the set adding amount and adding time; according to the invention, the addition amount of the complex microbial inoculant is dynamically calculated and adjusted, and accurate addition of the microbial inoculant is realized.

The invention discloses a high-arsenic hot spring replenishment surface water resource quantity determination method, device and equipment, a medium and a product, and relates to the technical field of water resource evaluation and ecological environment protection. The method comprises the following steps: acquiring information data of a target research area; screening symbolic characteristic parameters of the information data by adopting a statistical analysis method and geochemical diagrams to obtain screened information; based on the law of conservation of mass and the end member mixing theory, a multivariate mixing model is constructed according to the screening information; solving the multivariate mixed model by adopting a least square method to obtain a mixing ratio; the mixing ratio is used for analyzing spatial distribution characteristics of the mixing process of the high-arsenic hot spring and the surface water in combination with water flow path survey information; and determining the replenishment resource quantity of the high-arsenic hot spring according to the mixing proportion. The method aims at improving the accuracy and efficiency of determining the surface water resource quantity supplied by the high-arsenic hot spring.

The invention discloses a boiler intelligent combustion accurate air distribution dynamic balance method. The method comprises the steps of multi-dimensional data acquisition, data preprocessing and feature extraction, combustion state intelligent identification, air distribution parameter dynamic optimization and closed-loop control execution. According to the boiler intelligent combustion accurate air distribution dynamic balance method, fuel characteristics, combustion states, flue gas components and air distribution operation full-dimension information are comprehensively captured through a multi-dimension data acquisition system, a scientific feature extraction method is combined to provide a solid basis for combustion state recognition, the combustion type and the unbalance degree are accurately judged by means of a CNN-LSTM mixed model, and the intelligent combustion accurate air distribution dynamic balance method has the advantages of being high in practicability and high in practicability. Then, targeted optimization of the flow, the air speed and the spraying angle of primary air, secondary air and tertiary air is achieved through a dynamic balance algorithm, the problems that ignition is difficult due to insufficient primary air, combustion is insufficient due to uneven distribution of secondary air, and pollution is increased due to excessive tertiary air are solved from the source, air distribution parameters are highly matched with combustion working conditions, and the combustion efficiency is improved. And the combustion process is promoted to tend to an ideal equilibrium state.

The invention relates to the technical field of atmospheric environment measurement and control, and particularly discloses a carbon dioxide column concentration multi-source spatio-temporal data fusion method, which comprises the following steps: by combining a hybrid model of HGT and Transformer, reanalyzing meteorological variables such as modeled XCO2 and wind speed and direction of ERA-5 in a product by utilizing XCO2 and CAMS-IO inverted by an orbital carbon observation satellite OCO-2 satellite, enhancing a CAMS-EGG4 data set, and obtaining a CAMS-EGG4 data set; according to the method, multi-source input features are fused in spatio-temporal joint modeling to realize high precision and strong generalization ability, a heterogeneous spatio-temporal diagram is constructed by combining a spatial proximity relationship with time sequence nodes, so that an internal spatio-temporal dependency structure in atmospheric CO2 observation is explicitly expressed, and the spatial coverage breadth advantage can be maintained while the spatial coverage breadth advantage is maintained. And the observation precision comparable with that of foundation observation is realized.

The invention discloses a special equipment operator electroencephalogram fatigue detection method based on sample compensation and a hybrid model, and the method comprises the steps: collecting an electroencephalogram signal of a special equipment operator, and carrying out the preprocessing of the electroencephalogram signal, and obtaining a preprocessed electroencephalogram signal; performing multi-domain feature extraction and fusion on the preprocessed electroencephalogram signals to obtain a multi-dimensional feature set; performing compensation processing on the electroencephalogram signal and the multi-dimensional feature set by adopting a sample compensation method to obtain a compensated sample feature set; wherein the sample compensation method comprises a characteristic level compensation method and a signal level compensation method; training a hybrid model through the compensated sample feature set to obtain a trained hybrid model; wherein the hybrid model comprises a deep learning model and a traditional machine learning model; and performing electroencephalogram fatigue detection on the special equipment operator through the trained hybrid model to obtain a detection result.

A method and system for anomaly detection from time-series input data. A Gaussian Mixing Model (GMM) learns distribution parameters in an offline learning stage using sample data. The data used for offline learning and for a subsequent online anomaly detection stage are time-series data collected for multiple parameters of a machine operation, such as a robot performing a repetitive set of operations. The method includes aligning the data with a known good reference data file and taking a difference from it before providing the data to the GMM. In an online anomaly detection stage, the GMM calculates a probability that each time-series data point fits the distribution, and then a log-sum calculation is performed on each data file to determine the likelihood that the file contains anomaly data.The log likelihood of the file is compared with previous values, and an alert is issued if there are statistical deviations from the historical data.

The application discloses a data intelligent service method and device under a mixed model, equipment and a storage medium, relates to the technical field of artificial intelligence technology, and comprises the following steps: acquiring user question and answer information and user search information; identifying at least one of a corresponding query intention category, an inquiry intention category, an exploration behavior category and a search behavior category based on the user question and answer information or the user search information, and determining at least one of query intention category information, inquiry intention category information, exploration behavior category information and search behavior category information; predicting a corresponding service demand based on at least one of the query intention category information, the inquiry intention category information, the exploration behavior category information and the search behavior category information, and determining service demand prediction information; and analyzing a corresponding data resource based on the service demand prediction information, and determining a data resource analysis result. The application identifies query, inquiry, exploration and search intentions, performs data interpretation, and adapts intelligent service.

The application discloses a sandwich gold purity evaluation method and device based on a mixed model, equipment and a medium. The method comprises the following steps: collecting original data; inputting the processed original data into a mixed model for training to obtain a trained mixed model; wherein the mixed model comprises a feature extraction module, a weighted fusion unit and a classification module connected in sequence; collecting a second pulse eddy current signal and a second supplementary feature corresponding to the to-be-detected gold; inputting the second pulse eddy current signal and the second supplementary feature into the trained mixed model to output a prediction vector; converting the prediction vector into a one-hot code to obtain the number of the to-be-detected gold through the one-hot code. The problem that the existing classification mixed model cannot be applied to the high-fraud gold purity detection scene is solved, the classification accuracy of the fraud gold purity detection is improved, and the application range of the classification mixed model in the rare sample detection field is widened.

The invention discloses an iOS application background task scheduling optimization method and system, and the method comprises the steps: collecting perception data of an accelerometer and an ambient light sensor and network connection mode data, carrying out the analysis and processing of the data through a decision tree and Bayesian network mixed model, and carrying out the recognition of an application scene according to an analysis and processing result; starting a three-dimensional scoring model based on the obtained application scene recognition result, and performing comprehensive value quantification on background tasks in combination with a multi-dimensional dynamic evaluation system; and dynamically allocating CPU and network resources by adopting an elastic time window algorithm according to a comprehensive value quantification result, responding to system resource changes in real time by matching with a lightweight state synchronization mechanism, and scaling background task execution intensity in real time in combination with the residual electric quantity of the equipment and the current temperature state. The method solves the problems that the execution efficiency of the background task of the iOS equipment is low, the system resource distribution is rigid, the adaptability of the existing optimization scheme is poor, and the equipment state perception is insufficient.

The application relates to the computer technical field and discloses a modeling method of a statistical mixed model in a big data distributed scene. The method comprises the following steps: distributing and storing data shards; initializing model parameters; iteratively performing an expectation step and a maximization step; the expectation step is scheduled to a GPU node to perform parallel calculation on posterior probability; the maximization step is scheduled to a CPU node to aggregate statistics and update parameters, perform component merging / deletion, and perform convergence judgment; meanwhile, a memory reuse mechanism based on reference counting and scope analysis is adopted to reduce redundant data transmission. Through heterogeneous task scheduling and memory collaborative optimization, the training speed, resource utilization rate, and model self-adaptation capability are improved.

The invention belongs to the technical field of frequency spectrum prediction, and particularly relates to a frequency spectrum state prediction method based on a hybrid deep learning model, the hybrid deep learning model is fused with a long short-term memory (LSTM) network and a multi-layer perceptron (MLP), and through three-dimensional frequency spectrum data sensing, self-adaptive dual-threshold energy detection and hybrid model prediction, the frequency spectrum state is predicted. And the accuracy of idle channel spectrum prediction is further improved. According to the method, the secondary user (SU) in the cognitive radio system (CRS) can quickly select the channel with the highest idle probability for access, the repeated sensing frequency is reduced, the total sensing time is reduced by 30%, and the effective data transmission time is improved by 30%. Compared with the prior art, the method provided by the invention is higher in frequency spectrum state prediction precision in a low signal-to-noise ratio (SNR) scene, the throughput of the system is remarkably improved, and the energy consumption of the system is lower.

The invention provides a method for predicting pathogenicity of functional non-coding copy number variation in brain diseases, and belongs to the technical field of bioinformatics. Comprising the steps that brain-related features are collected to construct a brain-related feature set, and the types of elements in the brain-related feature set comprise transcription factor binding sites, histone modification and chromatin accessibility; constructing a hybrid model fusing two-dimensional convolution and swing-transformer for capturing collaborative regulation characteristics of brain-related cis-regulatory elements in the DNA sequence and generating N-dimensional brain-related functional feature annotations, and training the hybrid model by using a brain-related feature set; cNV data sets of the two pieces of brain-related copy number variation data are constructed, the data types in the mixed data set comprise coding and non-coding, and the high-confidence non-coding data set only comprises non-coding data; training a random forest model based on the two CNV data sets; the hybrid model and the random forest model jointly form a DeFunCNV model, and the pathogenicity of functional non-coding copy number variation in brain diseases is predicted.

The application discloses a view optimization query method and system for a multi-mode database, and relates to the field of computer view optimization query. A multi-mode JSON view set for a multi-mode database query scene is constructed; a multi-mode database system sequentially passes through a query language analysis module, a query language optimization module and a query language execution module according to the arrival order of a streaming mixed query load Qn to perform data processing; a mixed query request Q1 passes through the query language analysis module to generate a logical execution plan LP1; the LP1 passes through the query language optimization module to generate a physical execution plan PP1; the PP1 passes through the query language execution module to generate a query result D1; until the streaming mixed query load Qn is completely processed, and n query results are returned. The view format of the application can be used for accelerating database query for a single data model, and supports fusion representation of multiple data models, is used for accelerating mixed model query, and solves the problem of model information loss in the multi-mode view.

The invention discloses a carbapenem drug resistance marker screening method and system based on a cross-species compressed Debrueine diagram and a medium. The method comprises the following steps: starting from whole genome sequencing data of gram-negative bacteria belonging to different species and carbapenem drug phenotypes of the gram-negative bacteria, constructing a compressed Debrueine graph based on cross-species joint data, and taking existence / deletion of nodes in the graph as unified genetic variation characteristics. Performing correlation analysis on the nodes and the drug resistance phenotypes by using a linear hybrid model to obtain a candidate node set related to the phenotypes; k-mer is extracted based on the candidate node sequence, and secondary statistical screening is completed in combination with chi-square test and mutual information; and finally determining a group of carbapenem drug-resistant genetic markers which can be applicable across species through a hierarchical feature selection strategy of random forest and XGBoost. Efficient dimension reduction of large-scale cross-species genome data, cross-species consistent variation representation and high-interpretability marker screening are achieved.

The invention relates to a dynamic project cost intelligent prediction method, system and equipment, and the method comprises the steps: anchoring a prediction dimension through a cost motivation map through a closed-loop process of cost motivation quantification, multi-source data structuring, feature engineering and hybrid model dynamic prediction, and carrying out the directional mapping from multi-source data to a cost motivation dimension, performing corresponding feature processing according to feature types on the basis of feature classification defined by the cost motivation quantization atlas to form a structured feature set for model training; project data of different stages are input into a trained mixed cost prediction model, dynamic project cost prediction is achieved, and the mixed cost prediction model comprises an XGBoost model used for extracting static features and an LSTM model based on extracted historical time series data features. Compared with the prior art, the method has the advantages of realizing more accurate and more adaptive project full-stage dynamic cost prediction and the like.

The invention relates to the technical field of ice rink environment control, and particularly provides an ice rink anti-fog prediction method based on machine learning. Comprises: collecting historical environment parameters, and dividing into a training set and a test set; then, a logistic regression model is adopted to quickly calculate the fogging probability, and meanwhile, a GSA-LSTM-Transformer mixed model optimized through a golden sine algorithm is utilized to predict the change trend of environmental parameters; according to the model, the long-term time sequence feature capture capability of the LSTM and the global feature extraction advantage of the Transform are combined; meanwhile, a Levy-GWO optimization algorithm is introduced to carry out dynamic optimization on hyper-parameters of the Stacking integration model; according to the algorithm, the convergence precision and generalization ability of the model are remarkably improved by combining swarm intelligent search of a grey wolf optimization algorithm with a global disturbance strategy of Levy flight; according to the method, through multi-model fusion and intelligent optimization, accurate prediction and dynamic prevention and control of the fogging risk are realized.

The invention provides a Baijiu production method and system based on machine learning, and the method comprises the steps: constructing a multi-source heterogeneous data collaborative collection system, and collecting various data in a Baijiu production process; carrying out intelligent processing on the collected data; a hybrid model architecture and a transfer learning strategy are adopted, a dynamic adaptive model optimization mechanism is established, and the model performance is monitored in real time. Compared with the prior art, on the aspect of data, a multi-source heterogeneous acquisition system is constructed, and an advanced data management technology is adopted, so that the data quality and the value mining capability are improved; in the aspect of the model, the adaptability is enhanced by the hybrid architecture, the rapid customization of the model is realized, a dynamic adaptive optimization mechanism is established, and the continuous and efficient operation of the model is ensured; cost and risks are reduced in multiple dimensions such as hardware, software and a decision-making mechanism, and feasibility and reliability of technology application are enhanced.